This section is concerned with the version of K Nearest Neighbor which considers both the feature similarity and the feature value one. The sample words are encoded into numerical vectors whose features are texts by the scheme which was described in section 3.1. The novice word is given as the classification target, and it is also encoded into a numerical vector. Its similarities with the sample words are computed by equation (3) for selecting nearest neighbors, in the proposed version. Therefore, in order to provide the detail algorithm, we describe the proposed KNN version, together with the traditional one.